Figure. Vitreoretinal surgical techniques used in association with permanent Boston Type 1 keratoprosthesis (KPro) surgery: 3-port 25-gauge pars plana vitrectomy (PPV) with sutureless, self-sealing cannula sites (A); temporal approach for patient 3 (intraocular lens extraction [B] and 23-gauge PPV [C]); exploratory endoscopy as a prelude to KPro surgery in first operation for patient 9 (D); view of fundus through KPro (E); fundus showing pale and cupped nerve and choroids in patient 1 during first operation (F); and use of perfluoro-n-octane as medium-duration tamponade (G).
Kiang L, Sippel KC, Starr CE, Ciralsky J, Rosenblatt MI, Radcliffe NM, D’Amico DJ, Kiss S. Vitreoretinal Surgery in the Setting of Permanent Keratoprosthesis. Arch Ophthalmol. 2012;130(4):487-492. doi:10.1001/archophthalmol.2011.1115
Author Affiliations: Weill Cornell Medical College/Rockefeller University/Sloan-Kettering Institute Tri-Institutional MD-PhD Program (Dr Kiang) and Cornea (Drs Sippel, Starr, Ciralsky, and Rosenblatt), Glaucoma (Dr Radcliffe), and Retina (Drs D’Amico and Kiss) Services, Weill Cornell Medical College, New York, New York.
Objectives To evaluate the surgical management of vitreoretinal pathology in patients with a permanent Boston Type 1 keratoprosthesis (hereafter referred to as a KPro) in the era of small-gauge vitrectomy techniques.
Methods Retrospective review of 23 small-gauge vitreoretinal surgical procedures during or after Dohlman-Doane KPro placement in 14 eyes.
Results Established and innovative techniques were used, including sutureless small-gauge vitrectomy, temporal positioning of surgeon, long-term tamponades, and exploratory endoscopy. Retro-KPro membranes formed less frequently when vitrectomy was performed during KPro placement. Anatomical goals were achieved, and no serious complications directly resulted from these techniques. Visual acuity, frequently limited by preexisting pathology, improved in most cases.
Conclusions Modern posterior segment surgical techniques, including small-gauge sutureless vitrectomy, can be effectively used for patients with a permanent KPro. Vitrectomy and glaucoma tube revision by a team of subspecialists at the time of KPro placement may reduce subsequent complications.
Improved outcomes and expanding indications have increased the use of a permanent Boston Type 1 keratoprosthesis (hereafter referred to as a KPro) for visual rehabilitation.1,2 The increasing use of KPro, however, comes with the challenge of managing inevitable posterior segment complications in eyes that have undergone a KPro.
The last comprehensive overview of vitreoretinal techniques in this setting was published nearly a decade ago.3 Since that report, the preferences of vitreoretinal surgeons have shifted from 20-gauge vitrectomy to predominantly small-gauge sutureless procedures.4 With continued refinements in instrumentation, there are ever-expanding indications for small-gauge vitrectomy.5 Although management of endophthalmitis in eyes that have undergone KPro with 25-gauge vitrectomy was recently described, a comprehensive update on vitreoretinal surgical techniques with these implants has not been undertaken.6
The aim of our study was to describe and review the practicability and the anatomic and visual outcomes of, and the techniques used in, small-gauge vitrectomy in 23 cases performed on 14 eyes with a KPro. To the best of our knowledge, this is the largest series of eyes with KPro that have undergone small-gauge vitrectomy.
A retrospective review of 23 consecutive vitreoretinal procedures performed at Weill Cornell Medical College, New York, New York, in 14 eyes with a KPro was undertaken (Table and
Of 23 vitrectomies, 7 were performed at the time of KPro placement; 3 included a glaucoma procedure. Fifteen of the vitreoretinal procedures were in eyes with a preexisting KPro (including 1 KPro removal), 7 were performed at the time of KPro placement, and 1 was an exploratory endoscopy prior to KPro placement. Indications for KPro included corneal neovascularization, acid burn, corneal opacity, multiple graft failure, trauma, ulcers, and pseudophakic bullous keratopathy. Indications for vitrectomy included 8 cases with retro-KPro membranes (RPMs), with 7 of these cases occurring in eyes for which vitrectomy was not performed at the time of KPro placement; 6 cases of retinal detachment; 6 cases of vitreous debris or opacity; and 5 cases with epiretinal membranes.
Prior to surgery, a patient underwent an evaluation of his or her retina by use of indirect ophthalmoscopy (if possible) and high-resolution B-scan ultrasonography. In one case, exploratory endoscopy was performed to determine visual potential. When RPMs were present, an assessment was made to determine whether YAG capsulotomy would be effective or whether surgical intervention was necessary.
Of 23 cases, 22 underwent a 3-port small-gauge (23- or 25-gauge) vitrectomy (Figure, A); one case was via an open-sky approach. Excellent visualization of the posterior segment was routinely obtained using a 130° panoramic AVI contact lens (Figure, D and E); a Charles irrigating lens was used for macular maneuvers. The contact lens system provided a better visualization of the retina than did a noncontact lens system such as the BIOM 3 Non-Contact Panoramic Wide-Angle Viewing System (Oculus, Inc). Microcannulas were inserted from 1 to 3.5 mm posterior to the limbus. Infusion sites could then be changed, allowing conversion to a temporal surgical approach in one case (Figure, B). In conjunction with vitrectomy, an endolaser, silicone oil tamponades, perfluoro-n-octane (Figure, F), and intraocular steroids were used. Cataracts were removed using the small-gauge vitrectomy handpiece; intraocular lenses (IOLs) were taken out via an open-sky approach. Endoscopy was performed in one case (Figure, C). The majority of sclerotomy sites were not sutured: 13 of 15 cases of 25-gauge vitrectomy and 2 of 7 cases of 23-gauge vitrectomy were sutureless.
Of 23 vitrectomies, 15 were for posterior segment complications in eyes with a preexisting KPro, 4 were performed at the time of KPro placement in eyes with vitreous debris and retained a lens fragment or a subluxated posterior chamber IOL, 3 occurred at the time of KPro placement without posterior segment complications, and 1 was performed as a prelude to KPro placement for retinal evaluation. The indications for vitrectomy in these cases included: (1) lensectomy at the time of KPro placement; (2) glaucoma shunt revision with IOL removal; and (3) glaucoma shunt revision. The most common indication for pars plana vitrectomy (PPV) with a preexisting KPro was the presence of RPMs in 7 cases, followed by retinal detachment, vitreous opacity or debris, and the presence of epiretinal membranes. The presence of a single RPM occurred in a patient who underwent a vitrectomy at the time of KPro placement. Other concurrent pathologies included choroidal hemorrhage and proliferative vitreoretinopathy. Additionally, patient 2 presented with acute endophthalmitis 16 months after KPro placement. The mean follow-up period for all eyes was 11.3 months (median, 7 months; range, 1 day to 40 months).
In all 8 cases, the RPMs were successfully removed. Visual acuity improved in 3 patients7- 9 and remained unchanged in 5 patients, with the unchanged visual acuities being attributable to preexisting retinal or optic nerve pathology.
Of the 6 cases of retinal detachment that underwent PPV, 4 cases attained a complete anatomical reattachment. In patient 2, owing to retinal holes and fibrosis, the retina was judged to be beyond meaningful repair during the PPV; thus, complete anatomical reattachment was not pursued. Patient 9 had a suspected retinal detachment that was found to be a shallow macular detachment on examination using exploratory endoscopy. Visual acuity improved in 2 patients.7,8 Visual acuity remained unchanged in patient 2 because the retina was not completely reattached. Patient 5 had a total retinal detachment. Patient 1 had concurrent choroidal hemorrhage that limited our determination of the patient's final visual acuity.
Removal of vitreous opacities was successful in all 6 cases. Visual acuity improved from hand motions to 20/70 in patient 10, from 20/500 to 20/20 in patient 6, and from hand motions to 20/100 in patient 11.
For the cases in which PPV was performed at the time of KPro placement without posterior segment pathology, visual acuity improved from a range of light perception to 20/500 to counting fingers at 3 ft to 20/20. A significant amount of potentially proinflammatory anterior debris was removed along with the IOL.
In all 5 patients who underwent epiretinal membrane peeling, the membranes were successfully removed. In these patients, visual acuity changes, from a range of light perception to 20/200 before surgery to a range of light perception to 20/300 after surgery, were attributable to membrane peeling, the clearing of vitreous opacities, and KPro placement (Table).
Eyes with a KPro present a unique challenge for posterior segment surgery. Since the last report,3 indications for KPro placement have expanded, and vitreoretinal surgical techniques, including small-gauge vitrectomy, have advanced. We have successfully adapted these cutting-edge vitrectomy techniques to eyes that have underwent a KPro, with favorable outcomes. We describe the solutions to unique complications in these eyes and the strategies to prevent such complications.
In the original report,3 a modified 2- and 3-port 20-gauge vitrectomy was described. Herein, we illustrate trocar-based 23- and 25-gauge vitrectomy techniques in the setting of KPro. In our experience, vitrectomy in a closed globe, as opposed to a previously described open-sky approach,3 is most advantageous in minimizing the risk of choroidal hemorrhage during the surgery. The use of a trocar-based system permits effortless exchange of the infusion site, as exemplified with patient 3, for whom the temporal position was used for improved retinal visualization.7 Smaller incisions with 23- or 25-gauge systems may also be preferable given the extent of scarred conjunctiva and foreshortened fornices frequently found in these cases. Interestingly, even in these scarred eyes, the small-gauge incisions are oftentimes self-sealing and do not require a suture at the sclerotomy site. The use of this type of incision may also decrease the inflammatory response in an already inflamed eye.
The most frequent posterior segment complication encountered was the presence of RPMs, the incidence of which is reported to be up to 40%.2,8,10 The energy required to perform a YAG capsulotomy on these thick membranes may lead to a retinal detachment, a giant retinal tear, or persistent ocular inflammation. Such complications were noted in patient 4, who ultimately required a vitrectomy. Of the 15 cases with a preexisting KPro, 7 (47%) had RPMs. In the 7 eyes for which a vitrectomy was performed at the time of KPro placement, only a single RPM formed (14%). Removal of potentially proinflammatory material during KPro placement may decrease the incidence of RPMs. With IOLs, however, the presence of RPMs presents a special challenge. Although RPMs can be successfully removed from between the IOL and the KPro (patient 10), there is a risk of displacing the IOL or damaging the prosthesis. We advocate vitrectomy at the time of KPro placement because removing the IOL and clearing the inflammatory material may decrease the incidence of RPM formation and post-KPro inflammation.
The second most frequent posterior segment complication requiring PPV was retinal detachment. In the previous series,3 the authors alluded to the need for a tamponade technique. We now report experience with both perfluoro-n-octane and a silicone oil tamponade. Although not routinely recommended as a prophylaxis for retinal detachment in all cases of KPro, a silicone oil tamponade can be successfully used with a KPro (eg, for patients 1, 8, and 9). Perfluoro-n-octane may also be left in place to temporarily tamponade inferior retinal pathology (patient 5). However, given its inflammatory nature, short-term use of perfluoro-n-octane for 1 or 2 weeks requires concomitant topical and, at times, intravitreal steroid treatment.
Patients with a KPro may also develop epiretinal membranes that require peeling. As evidenced by patients 1, 7, 8, 10, and 12, anatomic and visual success can be achieved using small-gauge techniques and disposable instruments, even through a KPro. Standard visualization and surgical maneuvers can be applied in this setting without obvious complications related to a small-gauge PPV.
Vitreous biopsies were performed to aid in the management of several of cases. In patient 2, who developed endophthalmitis after being lost to follow-up for several months, a vitreous tap with the small-gauge infusion and vitrector provided a controlled means to enter a severely inflamed and scarred eye. In patient 4, who had recurrent uveitis after a YAG capsulotomy, and patient 6, who had recurrent non-Hodgkin lymphoma, a diagnostic small-gauge vitrectomy provided sufficient samples for treatment decisions. Small-gauge techniques are ideal in this setting because adequate specimen retrieval can be accomplished with minimal disruption to the eye.
By the nature of the disease process that results in the need for a KPro, the view of the fundus is typically extremely poor, and it may be difficult to evaluate visual potential. Although B-scan ultrasonography is useful, the best determination often comes with direct visualization. As such, an exploratory endoscopy (via a 20-gauge sclerotomy) was performed to evaluate whether visual potential would merit KPro placement in patient 9. The optic nerve, vessels, and macula were well visualized, and although a shallow macular detachment was noted, the tissues were judged to have adequate viability. This brief procedure allows for a more accurate assessment of potential visual recovery and of potential anatomical challenges prior to proceeding with KPro placement and may obviate the need for extensive surgery.
In our series, a vitrectomy was frequently performed at the time of KPro placement. The advantages of combining these 2 procedures (ie, PPV and KPro) are that, together, they (1) provide valuable information on the patient's visual potential (including a careful evaluation of retinal and optic nerve status), (2) remove potentially inflammatory material, and (3) prepare the eye for glaucoma drainage procedures during or subsequent to KPro placement.9
For similar reasons, we also favor the removal of preexisting IOLs at the time of KPro placement. First, potential inflammatory material around the IOL can be removed, which may reduce the incidence of posterior segment complications and RPM formation. In all 3 cases in which the IOL was removed at the time of KPro placement, substantial potentially inflammatory material was noted in the remnant capsular bag. This may have previously contributed to the repeatedly failed penetrating keratoplasties and may have also contributed to post-KPro inflammation. Second, should RPMs appear, it is challenging to remove them from between the prosthesis and the IOL, as in patient 10. Third, in any subsequent surgery, there is a risk of displacing the IOL. The absence of an IOL in an eye that has undergone a KPro allows the surgeon to access the back of the eye in a more controlled way. Finally, there are limitations to vitreoretinal maneuvers in eyes with IOLs (for instance, using a silicone oil tamponade with a silicone IOL). Although some anterior segment surgeons prefer to maintain an IOL/iris diaphragm as a structural aid to facilitate suturing of the graft during KPro placement, it is our opinion that the disadvantages of a pseudophakic KPro outweigh any purported advantages, especially given our combined team surgical approach and the simple incorporation of the additional optical power into the device itself.
Given the complexity of eyes that have undergone a KPro, we have taken a team approach to the management of these patients, with joint cases between corneal and retinal surgeons at the time of KPro placement. Small-gauge vitrectomy at the time of KPro allows for an optimal view of the retina, evaluation of visual potential, removal of IOL if desired, a vitreous biopsy, clearance of inflammatory material that may prevent RPM formation, and preparation for glaucoma surgery. In the setting of an aphakic unicameral eye with severe abnormal anterior segment anatomy, glaucoma drainage devices are typically placed more posterior (sulcus or pars plana) and are thus prone to blockage if an adequate PPV has not been performed. When indicated, we have also included a glaucoma specialist for revision of the frequently present glaucoma drainage devices. Cornea, retina, and glaucoma specialists then follow these complex patients together.
Differing techniques and approaches to posterior segment surgery, such as small-gauge vitrectomy, membrane peeling, vitreous biopsy, and the use of tamponades, can be successfully applied to eyes with a KPro. As advances continue in the treatment of complications in eyes that have undergone a KPro, we must also consider the prevention of common complications such as retro-KPro membrane formation and treatment of glaucoma. We suggest that a PPV at the time of KPro placement has the potential to decrease the level of RPM formation by removing inflammatory material as well as the scaffolding on which these membranes may grow. Direct visualization of the retina by use of endoscopy or during PPV provides valuable information on the status of the retina. For the best long-term outcome in these eyes that have undergone a KPro, we advocate coordinated operations by retina, cornea, and glaucoma surgeons when appropriate and further follow-up by this multidisciplinary team. Continuing advances in vitreoretinal technology and in surgical techniques will further improve visual outcomes for the rapidly increasing number of patients receiving a KPro.
Correspondence: Szilárd Kiss, MD, Retina Service, Weill Cornell Medical College, 1305 York Ave, 11th Floor, New York, NY 10021 (firstname.lastname@example.org).
Submitted for Publication: November 13, 2010; final revision received May 2, 2011; accepted May 27, 2011.
Author Contributions: All authors had full access to all the data in the study and take responsibility for the integrity and the accuracy of the data analysis.
Financial Disclosure: None reported.
Funding/Support: This work was supported in part by a departmental grant from Research to Prevent Blindness. Dr Kiang, a student of the Weill Cornell Medical College/Rockefeller University/Sloan-Kettering Institute Tri-Institutional MD-PhD Program, was supported by National Institutes of Health Medical Scientist Training Program grant GM07739.